Systems/Circuits

Deletion of Suppressor of Cytokine Signaling 3 fromForebrain Neurons Delays Infertility and Onset ofHypothalamic Leptin Resistance in Response to a HighCaloric Diet

Hayden J. L. McEwen, Megan A. Inglis, Janette H. Quennell, X David R. Grattan, and Greg M. AndersonCentre for Neuroendocrinology and Department of Anatomy, University of Otago, Dunedin 9054, New Zealand

The cellular processes that cause high caloric diet (HCD)-induced infertility are poorly understood but may involve upregulation of suppressorof cytokine signaling (SOCS-3) proteins that are associated with hypothalamic leptin resistance. Deletion of SOCS-3 from brain cells is known toprotect mice from diet-induced obesity, but the effects on HCD-induced infertility are unknown. We used neuron-specific SOCS3 knock-out miceto elucidate this and the effects on regional hypothalamic leptin resistance. As expected, male and female neuron-specific SOCS3 knock-out micewere protected from HCD-induced obesity. While female wild-type mice became infertile after 4 months of HCD feeding, infertility onset inknock-out females was delayed by 4 weeks. Similarly, knock-out mice had delayed leptin resistance development in the medial preoptic area andanteroventral periventricular nucleus, regions important for generation of the surge of GnRH and LH that induces ovulation. We therefore testedwhether the suppressive effects of HCD on the estradiol-induced GnRH/LH surge were overcome by neuron-specific SOCS3 knock-out. Althoughonly 20% of control HCD-mice experienced a preovulatory-like LH surge, LH surges could be induced in almost all neuron-specific SOCS3knock-out mice on this diet. In contrast to females, HCD-fed male mice did not exhibit any fertility decline compared with low caloric diet-fedmales despite their resistance to the satiety effects of leptin. These data show that deletion of SOCS3 delays the onset of leptin resistance andinfertility in HCD-fed female mice, but given continued HCD feeding this state does eventually occur, presumably in response to other mecha-nisms inhibiting leptin signal transduction.

Key words: GnRH; high caloric diet; high fat diet; leptin; SOCS3

IntroductionMammalian fertility is governed by the hypothalamicgonadotropin-releasing hormone (GnRH) neuronal network.

The activity of GnRH neurons is regulated by metabolic factors,in particular the adipocyte-derived hormone leptin, which is per-missive for fertility (Barash et al., 1996; Chehab et al., 1996; Chuaet al., 1996; Donato et al., 2011a). Although leptin acts centrally tomodulate pulsatile GnRH release (Lebrethon et al., 2000; Naga-tani et al., 2000), this occurs indirectly via neurons afferent toGnRH cells because they do not contain leptin receptors (Quen-nell et al., 2009). Potential upstream sites of action for leptin’seffects include the ventral premammillary nucleus (PMV), whereleptin receptors have been shown to be sufficient for fertility infemale mice (Donato et al., 2011b), the arcuate nucleus (ARC),where many of the leptin-responsive GABAergic neuronsthat have been shown to be required for normal fertility in mice

Received July 4, 2014; revised May 22, 2016; accepted May 27, 2016.Author contributions: J.H.Q., D.R.G., and G.M.A. designed research; H.J.L.M., M.A.I., J.H.Q., and G.M.A. performed

research; H.J.L.M. and G.M.A. analyzed data; G.M.A. wrote the paper.This work was supported by the Health Research Council of New Zealand and New Zealand Lottery Health

Research grants. We thank G. Schutz and W. Alexander for supplying transgenic mice.The authors declare no competing financial interests.Correspondence should be addressed to Dr. Greg M. Anderson, Centre for Neuroendocrinology and Department of

Anatomy, University of Otago, Dunedin 9054, New Zealand. E-mail: greg.anderson@otago.ac.nz.DOI:10.1523/JNEUROSCI.2714-14.2016

Copyright © 2016 the authors 0270-6474/16/367142-12$15.00/0

Significance Statement

Obesity is commonly associated with infertility in humans and other animals. Treatments for human infertility show a decreasedsuccess rate with increasing body mass index. A hallmark of obesity is an increase in circulating leptin levels; despite this, the brainresponds as if there were low levels of leptin, leading to increased appetite and suppressed fertility. Here we show that leptinresistant infertility is caused in part by the leptin signaling molecule SOCS3. Deletion of SOCS3 from brain neurons delays the onsetof diet-induced infertility.

7142 • The Journal of Neuroscience, July 6, 2016 • 36(27):7142–7153

(Zuure et al., 2013) are located, and the anteroventral periven-tricular nucleus (AVPV) and surrounding medial preoptic area(mPOA). Estradiol acts on neurons within the AVPV during theperiod immediately preceding ovulation to induce a switch fromnegative to positive estrogenic feedback (Le et al., 1997; Smith etal., 2006), causing a surge of GnRH and consequently luteinizinghormone (LH) from the anterior pituitary gland, followed byovulation (Homburg et al., 1976; Franz, 1988). Leptin may regu-late estrous cyclicity and LH release in mice via mPOA neuronalnitric oxide-producing cells (Bellefontaine et al., 2014).

Obesity is associated with a deregulation of the hypotha-lamo-pituitary-gonadal axis. While exogenous gonadotropinstimulation can improve fertility in chronically high caloricdiet (HCD)-fed mice (Tortoriello et al., 2004), treatments forhuman infertility show a decreased success rate with increas-ing body mass index (Dokras et al., 2006; Koning et al., 2010;Souter et al., 2011). A hallmark of obesity is an increase incirculating leptin concentration (van Dielen et al., 2002; Tor-toriello et al., 2004). Despite this, the brain responds as if therewere low levels of leptin, leading to increased appetite andsuppressed fertility, implying central leptin resistance (Wau-man and Tavernier, 2011). In animals models of diet-inducedobesity, leptin resistance can be quantified by the loss of asignaling response to exogenous leptin treatment (Munzberget al., 2004), but recent experiments using leptin receptor an-tagonists imply that endogenous leptin sensitivity may bemaintained (Ottaway et al., 2015).

Chronic HCD feeding leads to hypothalamic upregulation ofsuppressor of cytokine signaling 3 (SOCS3) (Munzberg et al.,2004). SOCS3 acts as a feedback inhibitor of leptin signaling viathe JAK (Janus kinase)/STAT (signal transducer and activatorof transcription) pathway as well as of insulin signaling by bind-ing to the insulin receptor and by degrading insulin receptorsubstrates (Howard and Flier, 2006). Brain-specific SOCS3 dele-tion protects mice from diet-induced obesity (Mori et al., 2004),but the effects on HCD-induced infertility have yet to be estab-lished. We used neuron-specific SOCS3 knock-out mice, ex-pressed on a background strain prone to diet-induced infertility,to elucidate whether SOCS3 mediates the effects of high dietaryfat on the onset of infertility and regional hypothalamic leptinresistance.

Materials and MethodsAnimals and diets. Wild-type and transgenic mice (DBA/2J backgroundstrain) were obtained from the University of Otago animal breedingfacility and housed under conditions of controlled lighting (lights onfrom 0600 to 1800 h) and temperature (22 � 1°C). Except for overnightfasting as described, they had free access from the date of weaning toeither standard low fat and sucrose (4.8% fat, 0% sucrose by weight,Specialty Feeds; designated LCD) or high fat and sucrose (23% fat, 42%sucrose by weight, SF03– 020, Specialty Feeds; designated HCD) dietsthat were matched for protein, fiber, vitamin, and mineral content. TheUniversity of Otago Animal Ethics Committee approved all animal ex-perimental protocols.

CamKinaseII� Cre and SOCS3-flox transgenic mice were backcrossedonto a DBA/2J inbred mouse strain five times from a C57BL/6J strainbecause pilot studies in our laboratory showed that female DBA/2J micesuffer a 90% increase in the duration between litters in response tochronic HCD feeding compared with LCD-fed littermates, whereas HCDis without effect on the fertility of C57BL/6J mice. Similar findings havebeen reported by others (Tortoriello et al., 2004). Neuron-specificSOCS3 knock-out mice were bred as follows. Homozygous female Socs3flox mice (Socs3fl/fl; loxP sites flanking Socs3 coding exon 2) (Croker et al.,2003) were bred to male CamKII�-iCre BAC (referred to as CamKII�-Cre) mice (Casanova et al., 2001). CamKII� is expressed around birth in

almost all forebrain neurons but not in glial cells (Ouimet et al., 1984;Burgin et al., 1990). The resulting female Socs3fl/�,CamKII�-Cre micewere backcrossed to male Socs3fl/fl mice to generate Socs3fl/fl,CamKII�-Cre conditional knock-out mice (referred to as neuron-specific SOCS3knock-out mice). Socs3fl/fl littermates served as controls. Transgenic micewere identified by PCR analysis of genomic DNA isolated from tail biop-sies using the following primer sets, at an annealing temperature of 62°C.For CamKII�-Cre identification: GGT TCT CCG TTT GCA CTC AGGA, CCT GTT GTT CAG CTT GCA CCA G and CTG CAT GCA CGGGAC AGC TCT (290 bp product indicates the wild-type gene and 345 bpproduct indicates the Cre-expressing gene), for Socs3fl identification:TCT TGT GTC TCT CCC CAT CC and TGA CGC TCA ACG TGA AGAAG (�650 bp product indicates the wild-type gene and �800 bp productindicates the floxed gene), and to detect Socs excision: TCT TGT GTCTCT CCC CAT CC and ACG TCT GTG ATG CTT TGC TG (�375 bpproduct indicates excised gene). PCR analysis of experimental versuscontrol tissues showed that Cre-mediated flox excision of the Socs3 genewas specific to Cre-expressing mice and to brain tissue (Fig. 1). All micewere weighed every 2 weeks, except for female mice when they werepaired with males for fertility analysis at 75 d of age.

Characterization of HCD-induced leptin resistance in DBA/2J mice. Tocharacterize the development of obesity in the DBA/2J mouse strain, 6male and female wild-type DBA/2J mice, raised on a LCD until 12 weeksof age, were fed either an LCD or HCD for a further 12 weeks. Energyconsumption and body weights were monitored. At 9 –10 weeks on theexperimental diets, mice were fasted for 12 h and then subjected to eithera leptin (1 mg/kg i.p. recombinant mouse leptin; National Hormone andPituitary Program) or vehicle (0.1 M PBS) challenge at 1700 h. The re-spective diets were returned 30 min after the start of the dark phase(1830 h). Food intake was measured at 3 h after return of food (2130 h).A 7 d recovery period was given between vehicle and leptin challenges.After 12 weeks on the experimental diets, mice were fasted overnight andterminal trunk blood samples were collected for measurement of serumleptin concentration. Inguinal, retroperitoneal, and gonadal fat padswere collected and weighed.

Effect of neuron-specific SOCS3 deletion on puberty onset, estrous cyclic-ity, and fertility in LCD- and HCD-fed male and female mice. Femaleneuron-specific SOCS3 knock-out and control mice were weaned ontoeither an LCD or HCD (n � 7 or 8 per group) and monitored for pubertyonset by daily examination for vaginal opening and subsequently firstestrus, based on vaginal cytology. At 60 d of age, estrous cyclicity wasmonitored for 10 consecutive days by vaginal cytology. Between 75 and200 d of age, they were paired with wild-type C57BL/6J male mice forfertility analysis, maintaining their respective diets. The use of a mousestrain known to remain fully fertile on an HCD (i.e., C57BL/6J) as themating partner in all mating studies enabled any effects of diet to belinked to the experimental (DBA/2J) mice (G. M. Anderson, unpub-lished data; Tortoriello et al., 2004). Cages were checked daily for thepresence of litters, the date and size of the litter recorded, and the pupswere immediately removed and culled. Estrous cyclicity was again mon-itored for 10 d at the end of this period of fertility testing (215-d-old, after195 d on the diets).

Figure 1. Representative examples of agarose gel bands indicating brain-specific Socs3deletion in neuron-specific SOCS3 knock-out mice (right 3 lanes), but not control mice (left3 lanes). Middle lane contains a 50 base pair ladder.

McEwen et al. • SOCS3 Deletion Rescues from Leptin-Resistant Infertility J. Neurosci., July 6, 2016 • 36(27):7142–7153 • 7143

Male neuron-specific SOCS3 knock-out and control mice (n � 7 pergroup) were paired with adult wild-type C57BL/6J female mice between35 and 180 d of age; cages were checked daily for litters as above. Date ofpuberty was measured as date of first successful mating, determined bysubtracting 20 d of gestation from the date of birth of the first litter. Micewere weighed every 2 weeks; for females, a separate cohort was used tomonitor body weight to avoid the confounding effects of pregnancy.Weight at puberty was calculated using the measured body weights eitherside of the date of puberty, assuming a linear growth rate. For glucosetolerance tests, fasted (8 h) tail tip blood glucose concentrations weremeasured and animals then injected with a bolus of 1.5 g/kg intraperito-neal glucose in 100 �l physiological saline. Blood glucose concentrationswere measured at 15, 30, 45, 60, and 90 min thereafter using an automaticglucose monitor (Roche; Accu-Chek Performa).

Effect of neuron-specific SOCS3 deletion on the ability of estradiol toinduce a preovulatory-like GnRH/LH surge in LCD- and HCD-fed femalemice. Female neuron-specific SOCS3 knock-out and control mice werefed an HCD from weaning for 110 d (n � 5 or 6 per group) or 225 d (n �7–10 per group). They were then subjected to a preovulatory LH surgeinduction model as described previously (Wintermantel et al., 2006;Zuure et al., 2013). Briefly, mice were ovariectomized and had anestradiol-filled silicone rubber capsule, 1.0 mm internal, 2.1 mm externaldiameter (Dow Corning), filled with medical-grade silicone rubber ad-hesive (Dow Corning) containing 17-� estradiol (0.1 mg/ml adhesive,Sigma-Aldrich) implanted subcutaneously. Each mouse was given a1-cm-long capsule (containing 1 �g estradiol) per 20 g body weight. Sixdays after ovariectomy, mice received a subcutaneous injection of estra-diol benzoate (1 �g/20 g body weight, Sigma-Aldrich) at 0900 h to inducean LH surge. At 1800 h on the following day, the surge peak time (Win-termantel et al., 2006), a blood sample was collected via heart punctureand the mice were perfused via the heart with 20 ml of 4% PFA in PBS, pH7.4. The serum was harvested and stored at �20°C for LH radioimmu-noassay. Brains were collected and cryoprotected in 30% sucrose for latersectioning and dual immunohistochemistry for cFos and GnRH (110 dgroup only).

Effect of neuron-specific SOCS3 deletion on hypothalamic leptin sensitiv-ity in LCD- and HCD-fed female mice. Leptin-induced phosphorylationof signal transducer and activator of transcription (pSTAT3) was used toassess hypothalamic leptin sensitivity. Female neuron-specific SOCS3knock-out and control mice were fed either a LCD or HCD from weaningand subjected to a leptin challenge to determine hypothalamic leptinresponsiveness after either 110 or 225 d on the diets (n � 5–10 mice pergroup). These times were chosen based on the times that female controland neuron-specific SOCS3 knock-out mice became infertile in responseto the HCD in the present study. All mice were fasted overnight to reducethe concentration of endogenous circulating leptin. Recombinant mouseleptin (1 mg/kg) (National Hormone and Peptide Program) was injectedintraperitoneally, and mice were perfused 2 h later via the heart with 20ml of 4% PFA. Brains were collected and cryoprotected in 30% sucrosefor later sectioning and immunohistochemistry for pSTAT3.

Immunohistochemical staining for pSTAT3, GnRH, and cFos. Coronal(30-�m-thick) sections throughout the septal-preoptic area (sectionscollected for GnRH�cFos dual immunohistochemistry) or the entirehypothalamus (sections collected for pSTAT3 immunohistochemistry)were cut on a sliding microtome with a freezing stage to provide three setsof consecutive sections each (90 �m apart). The primary antibodies usedwere as follows: monoclonal rabbit anti-pSTAT3 (Tyr705, D3A7 XP; CellSignaling Technology), polyclonal rabbit anti cFos (AB5; Calbiochem),and polyclonal guinea pig anti GnRH antibody (GA02) (Rizwan et al.,2012). Omission of any of the primary antibodies resulted in a completeabsence of staining.

Before staining, sections were incubated for 30 min in 1% H2O2 and40% methanol to quench endogenous peroxidase activity. For pSTAT3staining, the sections were then incubated in 1 mM EDTA, pH 8.0, at 90°Cfor 15 min to unmask the antigen. The rabbit anti-pSTAT3 (1:2000 di-lution) or cFos (1:20,000 dilution) was diluted in a blocking solution of0.05 M TBS, 0.25% BSA, 2% normal rabbit serum, and 0.1% Triton X-100and incubated for 48 h at 4 C. This was followed by a 60 min incubationin biotinylated goat anti-rabbit IgGs (1:1000; Vector Laboratories), a 60

min incubation in Vector Elite ABC solution (1:200, Vector Laborato-ries), and a brief incubation in nickel-enhanced (0.8% Ni(NH4)2SO4)diaminobenzidine solution (Sigma-Aldrich) until blue-black nuclearstaining was clearly visible. For GnRH and cFos dual immunohistochem-istry, the tissues were next incubated for 48 h in guinea pig anti-GnRHprimary antibody (1:3000 dilution in the blocking solution describedabove), then for 60 min in HRP-conjugated rabbit anti-guinea pig IgGs(1:500; Dako) and then briefly in unenhanced diaminobenzidine solu-tion until a light brown cytoplasmic staining was visible. For dilution ofsecondary antibodies and washing the sections between each step, 0.5 M

TBS was used. Sections were mounted on (3-aminopropyl)-trithoxyl-silane-coated slides, dried, hydrated in water, dehydrated in graded seriesof alcohols, and then cleared in xylene before coverslipping with dibu-tylphthalate polystyrene xylene. Staining was observed using an OlympusAX70 Provis light microscope at 100� and 200� magnification.

GnRH neurons were counted as cFos-positive if they had a distinctblue-black-stained nucleus surrounded by brown cytoplasmic staining.All GnRH neurons in three sections per animal, which included themedial septum and region around the organum vasculosum of the lam-ina terminalis (0.4 – 0.7 mm anterior to bregma), as shown by Franklinand Paxinos (2008), were counted by an operator blind to the treatmentgroups. For pSTAT3, numbers of immunoreactive cells were countedusing the image analysis software ImageJ (National Institutes of Health)to identify and count darkly stained nuclei within uniformly sized hypo-thalamic regions and the Franklin and Paxinos (2008) mouse brain atlasto define these regions. The regions analyzed and their coordinates ante-rior to bregma were the organum vasculosum of the lamina terminalis(OVLT, 0.4 – 0.5 mm), the mPOA and the AVPV (both 0.3– 0.4 mm), theARC (�1.5 to �2.0 mm), and the PMV (�2.5 to �2.6 mm). Images werephotographed at 100� magnification. For each region, two or three sec-tions were counted and averaged to provide a single data point (countsper region per section) for each animal.

Blood serum immunoassay for leptin, LH, estradiol, and testosteroneconcentrations. Serum leptin concentration was assayed in 10 �l aliquotsusing a commercial ELISA kit (Millipore EZML-82K), according to themanufacturer’s instructions. The sensitivity of the assay (95% CI at 0ng/ml on the standard curve) was 0.1 ng/ml. For a serum pool falling inthe middle of the standard curve, the intra-assay CV was 8%.

Serum LH concentration was measured in 25 �l aliquots by radioim-munoassay. Values are expressed in terms of the rat standard NIDDK-ratLH-RP-3. Highly purified iodinated hormone (NIDDK-rat LH-I-10)was used as tracer, and the primary antiserum was NIDDK-rabbit anti-rat LH-S11 (final dilution 1:500,000). The sensitivity of the assay was 0.05ng/ml. For a serum pool falling in the middle of the standard curve, theintra-assay CV was 9%.

Serum total estradiol concentration was assayed in 25 �l aliquotsusing a commercial ELISA kit (Calbiotech, Mouse/Rat EstradiolELISA ES180S-100), according to the manufacturer’s instructions.The sensitivity of the assay was 2.5 pg/ml, and the intra-assay CV was5%. Serum total testosterone concentration was assayed in 10 �laliquots using a commercial ELISA kit (LDN Labor Diagnostika Tes-tosterone Rat/Mouse ELISA AR E-8000), according to the manufac-turer’s instructions. The sensitivity of the assay was 10 pg/ml, and theintra-assay CV was 11%.

Data analysis. All 2 � 2 factorial data were analyzed using two-wayANOVA with diet and genotype or diet and leptin/vehicle treatment asthe factors for comparison. Post hoc analysis was performed using theHolm–Sidak multiple-comparisons test. A preovulatory-like LH surgewas determined to have occurred if the plasma LH concentration was �2ng/ml. The proportion of mice exhibiting a preovulatory-like LH surgewas analyzed using a � 2 test. All other data were analyzed using Student’st tests or Mann–Whitney U tests. Results are presented as mean � SEM,and differences were considered significant at p 0.05.

ResultsCharacterization of HCD-induced leptin resistance inDBA/2J miceMale and female wild-type DBA/2J mice fed an HCD for 10 weekshad a 33%–50% higher caloric intake compared with LCD-fed

7144 • J. Neurosci., July 6, 2016 • 36(27):7142–7153 McEwen et al. • SOCS3 Deletion Rescues from Leptin-Resistant Infertility

controls (t(10) � 5.81 [male] and 6.71 [female], p 0.001), de-spite an 8%–16% reduced food intake on a mass basis (Fig. 2A).When fasted for 24 h and treated with 1 mg/kg leptin, LCD-fedmale and female mice consumed significantly less food over thefirst 3 h of refeeding compared with saline-treated controls (t(4) �5.49 [male] and 4.42 [female], p 0.01), whereas HCD-fed miceshowed no anorexigenic response to leptin (Fig. 2B), indicatingthe development of central leptin resistance in both sexes. In maleHCD-fed mice serum leptin concentration (Fig. 2C) and go-nadal, inguinal, and retroperitoneal fat pad masses (Fig. 2D) wereelevated 3- to 10-fold compared with LCD-fed controls (t(10) �10.53 [gonadal], 3.53 [inguinal], and 9.20 [retroperitoneal], p 0.005); in females, the same responses of fat mass to HCD feedingoccurred but were less pronounced (t(10) � 3.40 [gonadal], 2.87[inguinal], and 7.22 [retroperitoneal], p 0.01) (Fig. 2C,D).

In male (Fig. 3A) and female (Fig. 3B) control mice, the HCDsignificantly increased body weight; these mice were 30%– 65%heavier than LCD-fed mice from �6 weeks of age. As expectedbased on previously published data (Mori et al., 2004), neuron-specific SOCS3 knock-out completely protected the male micefrom HCD-induced obesity compared with their control litter-mates (two-way repeated-measures ANOVA on body weight:F(1,96) � 42.38, p 0.001). Post hoc analysis revealed that theHCD-fed control male mice were significantly heavier than theirSOCS3 knock-out counterparts from 15 weeks of age, respec-tively (Fig. 3A). The effect of SOCS3 knock-out was far less pro-nounced in females; although there was a significantly loweroverall body weight in the HCD-fed knock-out mice comparedwith the HCD-fed control mice (main group effect: F(1,15) � 5.09,p 0.05), post hoc analysis at specific time points revealed fewsignificant differences (Fig. 3A). Glucose tolerance was tested af-ter 120 d of HCD treatment. Female HCD-fed SOCS3 knock-outand control mice both exhibited impaired glucose clearance inresponse to a 1.5 g/kg intraperitoneal glucose challenge com-pared with their LCD-fed knock-out and control counterparts(35% higher area under the glucose response curve; two-wayANOVA main effect of diet: F(1,28) � 4.86, p 0.05). After 225 dof HCD treatment, total abdominal fat mass was similarly ele-vated in females HCD-fed SOCS3 knock-out and control mice(2.72 � 0.44 and 2.57 � 0.29 g, respectively) compared with theirLCD-fed knock-out and control counterparts (0.10 � 0.03 and0.14 � 0.14 g; two-way ANOVA main effect of diet: F(1,28) �87.05, p 0.001).

In males, HCD feeding also advanced the age of first matingrelative to LCD-fed mice, regardless of genotype (two-wayANOVA: F(1,28) � 18.28, p 0.001). Because the HCD-fedneuron-specific SOCS3 knock-out mice had a normal bodyweight, this meant that they became fertile at a lower weightthan all other mice (Fig. 3C; two-way ANOVA interaction ofbody weight and diet: F(1,28) � 4.31, p 0.05). In females,there was no significant effect of diet (which had only beeninitiated �10 d previously) or of genotype on age of firstestrus; however, the body weight at which this event occurredwas lower (by �2 g) in neuron-specific SOCS3 knock-outfemales fed either diet compared with controls (Fig. 3D; two-way ANOVA: F(1,28) � 4.31, p 0.01).

Adult fecundity in males was not affected by either HCD orgenotype. All males remained able to sire litters at regular inter-vals of �28 d (number of litters born over 145 d of mating: LCDcontrol, 4.3 � 0.5; LCD SOCS3 KO, 4.7 � 0.2; HCD control,5.0 � 0.4; HCD SOCS3 KO, 4.8 � 0.5) until the end of theexperiment (age when last litter was born: LCD control, 147 �8 d; LCD SOCS3 KO, 165 � 7 d; HCD control, 160 � 9 d; HCD

Figure 2. Characterization of the effects of high (23%) versus low (5%) fat diets on energyconsumption, leptin resistance, and adiposity in male and female DBA/2J mice. A, Daily energyconsumption (right bars), calculated over a 2 week period, was significantly elevated despite areduced food intake by mass (left bars). B, In 12 h fasted LCD-fed mice, leptin (1 mg/kg i.p.)significantly reduced 3 h food intake; whereas in HCD-fed mice, there was no response to leptin,indicating that they had become resistant to this hormone. Serum leptin concentration (C) andregional (gonadal, inguinal, and retroperitoneal) fat pad masses (D) were markedly increased inmale, and to a lesser extent in female, mice in response to the HCD. Mice (n � 6 per group) were5– 6 months old and were fed the treatment diets for 12 weeks. *p 0.05 versus LCD controls.**p 0.01 versus LCD controls.

McEwen et al. • SOCS3 Deletion Rescues from Leptin-Resistant Infertility J. Neurosci., July 6, 2016 • 36(27):7142–7153 • 7145

SOCS3 KO, 159 � 9 d). In marked contrast, the HCD markedlyattenuated the period of adult fertility in female mice (two-wayANOVA for age at birth of last litter: F(1,17) � 56.15, p 0.001;number of litters delivered: F(1,17) � 77.87, p 0.001). WhileLCD-fed females of both genotypes remained fertile until the endof the experiment (5.9 litters born over the mating study; averageinterlitter interval 22 d), HCD-fed control mice averaged only 2.2litters per dam and gave birth to their final litter at an average of45 d into the mating study. In contrast to this, neuron-specificSOCS3 knock-out prevented the abrupt loss of fertility at thistime in HCD-fed mice. The rescue from HCD-induced infertilityin response to SOCS3 disruption was limited to �4 weeks dura-tion, however; following 84 d of mating and having given birth to3.8 litters on average, neuron-specific SOCS3 knock-out micealso succumbed to HCD-induced fertility (Fig. 4A–C), signifi-cantly later than their control counterparts but earlier than theLCD-fed groups (two-way ANOVA interaction of genotype anddiet for age at birth of last litter: F(1,17) � 5.26, p 0.05; interac-tion of genotype and diet for number of litters delivered: F(1,17) �6.09, p 0.05). The eventual loss of fertility in both HCD-fedgroups was reflected in the pattern of estrous cycles at the begin-ning and end of the study. While mice in all four groups exhibitedregular cycles at 8 weeks of age as evidenced by their vaginalcytology, at the end of the study the frequency of estrous smears

was significantly reduced (from 28% at 8 weeks to 12% at30 weeks) in both the HCD-fed groups (two-way ANOVA:F(1,17) � 20.10, p 0.001), and diestrous-like cytological patternspredominated instead (data not shown). There were no diet orgenotype effects on plasma estradiol and total testosterone con-centrations after 120 d of HCD treatment in female mice.

To test whether the HCD was able to inhibit the estradiol-induced preovulatory GnRH/LH surge via a SOCS3-dependentmechanism, we next subjected additional neuron-specific SOCS3knock-out mice and control littermates, fed the HCD from wean-ing for 110 d, to an exogenous estradiol treatment protocol thatreliably causes activation of GnRH neurons and a preovulatory-like surge of LH (Wintermantel et al., 2006; Zuure et al., 2013).This age and duration of HCD exposure were chosen because itfalls within the window of time identified in the fertility experi-ment when neuron-specific SOCS3 knock-out mice are pro-tected from HCD-induced infertility (Fig. 4A). Only 20% ofcontrol HCD-fed mice experienced a preovulatory-like LH surge(serum LH concentration �2.0 ng/ml), implying that the HCDsuppressed the neuroendocrine control of ovulation at this time.In marked contrast, robust LH surges could be induced in almostall (83%) of neuron-specific SOCS3 knock-out mice on this diet(see Fig. 6A; �1,N�11�

2 � 11) � 4.41, p 0.05; mean serum LH

Figure 3. Effects of HCD and forebrain neuronal SOCS3 deletion growth profiles and onset of reproductive maturity. In male (A, n � 7 per group) and female (B, n � 7– 8 per group) mice, HCDsignificantly increased body weight gain, and this effect was prevented by neuronal SOCS3 deletion. Insets, Age of first successful mating (A, males) and vaginal opening (B, females). HCD feedingadvanced the date of first mating relative to LCD-fed mice regardless of genotype. The body weight at which the first fertile mating in HCD-fed males (C) and the first estrus in females fed either diet(D) occurred was advanced in neuron-specific SOCS3 knock-out mice compared with controls. *p 0.05 versus LCD controls. **p 0.01 versus LCD controls.

7146 • J. Neurosci., July 6, 2016 • 36(27):7142–7153 McEwen et al. • SOCS3 Deletion Rescues from Leptin-Resistant Infertility

concentration: U(9) � 4, p 0.05). There was a nonsignificanttrend (p � 0.095) for an increased percentage of GnRH neuronsactivated by estradiol (i.e., coespressing cFos) in the neuron-specific SOCS3 knock-out mice (Fig. 5A; for representative ex-amples for GnRH and cFos immunreactivity, see Fig. 5B). Whenthe ability of this estradiol treatment to induce a preovulatory-like LH surge was tested after 225 d of HCD feeding, when allmice in the fecundity test had become infertile, both control andSOCS3 KO mice exhibited low plasma LH concentrations (mean1.8 ng/ml, and 20% surging) (Fig. 5C). In a group of similarly

aged LCD-fed mice tested at the same time, 75% of mice surgedwith mean LH levels of 5.8 ng/ml.

Finally, we questioned whether these effects of HCD to causeinfertility and their prevention by SOCS3 knock-out were asso-ciated with altered leptin responsiveness in physiological systemsand in specific hypothalamic nuclei. For example, the AVPV is aregion known to be critical for generation of the preovulatoryGnRH/LH surge in mice surge (Le et al., 2001). Leptin acts in thisnucleus and in the surrounding medial preoptic area (Quennellet al., 2009; Scott et al., 2009) and, given the effects of HCD and

Figure 4. A, Timelines showing fecundity over 4 months for individual neuron-specific SOCS3 knock-out and control female mice fed either an LCD or HCD and paired with wild-type mates.Triangles represent birth of individual litters. Bold vertical lines indicate the average day that the last litter was born for each group. Shaded region represents the period for which neuron-specificSOCS3 knock-out was able to rescue mice from HCD-induced infertility. B, Comparison of the total number of litters born. C, Mean age at which the last litter was born (n � 4 – 6 per group).Recordings ceased after 180 d on the experimental diets (200 d of age). *p 0.05 versus control females.

McEwen et al. • SOCS3 Deletion Rescues from Leptin-Resistant Infertility J. Neurosci., July 6, 2016 • 36(27):7142–7153 • 7147

SOCS3 knock-out on surge generation described above, might beexpected to show altered leptin signaling in response to thesetreatments. Indeed, at 110 d of HCD feeding during the previ-ously identified window of time when neuron-specific SOCS3

knock-out mice are protected from HCD-induced infertility, sig-nificantly reduced leptin-induced pSTAT3 (indicative of leptinresistance) was observed in the mPOA and AVPV of femaleHCD-fed control mice compared with LCD-fed mice, whereasleptin responsiveness was maintained at normal levels in neuron-specific SOCS3 knock-out mice in these regions (Fig. 6B,C; two-way ANOVA interaction of genotype and diet for pSTAT3 inmPOA: F(1,24) � 6.71, p 0.01; interaction of genotype and dietfor pSTAT3 in AVPV: F(1,22) � 4.32, p 0.05). The same effect,albeit at a lower magnitude, was observed in the ARC (Fig. 6D;two-way ANOVA interaction of genotype and diet: F(1,28) � 6.43,p 0.01). In the OVLT and PMV, leptin sensitivity was unaf-fected by HCD or SOCS3 knock-out at this time (Fig. 6A,E). At 8months of age, however, leptin-induced pSTAT3 immunoreac-tivity was reduced in all HCD-fed in the OVLT, AVPV, and ARCrelative to LCD mice regardless of genotype (Fig. 6F,H,I; two-way ANOVA for pSTAT3 in OVLT: F(1,17) � 13.16, p 0.01;two-way ANOVA for pSTAT3 in AVPV: F(1,17) � 55.6, p 0.001;two-way ANOVA for pSTAT3 in ARC: F(1,17) � 105.1, p 0.001), and in the mPOA even LCD mice displayed low pSTAT3levels at this time (Fig. 6G). Only the PMV maintained a robustresponse to leptin in all groups, regardless of diet or treatment, atthis age (Fig. 6J). Representative examples of pSTAT3 immuno-staining depicting the regional loss of leptin responsiveness inneuron-specific SOCS3 knock-out mice between 110 and 225 dof HCD treatment are shown in Figure 6K–O.

Leptin responsiveness was also measured after 160 d of HCDin two systems known to be affected by this hormone: suppres-sion of voluntary food intake and stimulation of circulating LHlevels. As expected, fasted female LCD-fed mice showed a robustanorexigenic response to 1 mg/kg leptin, administered as de-scribed earlier (two-way ANOVA main effect of leptin treatment:F(1,18) � 8.94, p 0.01). This response was completely sup-pressed in HFD-fed mice, regardless of genotype (Fig. 7A), sug-gesting onset of leptin resistance in appetite-modulation circuitryat this stage of the dietary regimen regardless of SOCS3 KO. Asimilar pattern was seen for leptin stimulation of LH secretion(measured under the same conditions as the food intake experi-ment). Although 1 mg/kg leptin-treated LCD-fed mice inducedan increase in plasma LH concentration in tail tip blood after 20min compared with saline-treated mice (two-way ANOVA maineffect of leptin treatment: F(1,16) � 8.61, p 0.01), this responsewas not seen in either HFD-fed group (Fig. 7B).

DiscussionAlthough the correlation between obesity and infertility is widelyrecognized (Koning et al., 2010), the physiological causes ofHCD-induced infertility are poorly understood. Central resis-tance to leptin, a hormone essential for fertility, is known toaccompany this condition and is likely to be a major contributingfactor. Upregulation of SOCS3 when fed an HCD is one mecha-nism proposed to cause central resistance to leptin (Bjørbaek etal., 1998; Munzberg et al., 2004) and insulin (Howard and Flier,2006), of which the former is most important for reproductiveactivity (Quennell et al., 2009; Evans et al., 2015). Central dele-tion of SOCS3 protects against the development of obesity (Moriet al., 2004), but until now it has not been determined whetherthis is also the case in regards to fertility. Here we show thatSOCS3 deletion is protective against HCD-induced infertilityand leptin resistance in the mPOA, AVPV, and ARC of femalemice but becomes ineffective in this regard if the diet is main-tained for several months.

Figure 5. Neuron-specific SOCS3 knock-out prevents HCD-induced suppression of the estradiol-inducedLHsurgeinfemalemice.A,SerumLHconcentrationatthetimeofsurgeoccurrence(leftbars),percentage of mice exhibiting an LH surges (serum concentration �2 ng/ml) (middle bars), andpercentage of activated GnRH neurons (i.e., coexpressing cFos) (right bars) after 120 d of HCD feeding(n � 5 or 6 per group). B, Representative example of GnRH neurons (brown unenhanced DAB cyto-solic staining) that are either negative or positive (indicated by arrowhead) for cFos immunoreactivity(black nickel-enhanced DAB nuclear staining). Control and SOCS3 knock-out mice were fed an HCDfrom weaning for 110 d (sufficient to induce infertility in controls but not knock-outs). They were thenovariectomized and subjected to an estradiol-induced preovulatory surge induction model. C, Whenthe ability of this estradiol to induce an LH surge was tested after 225 d of HCD feeding, both controland SOCS3 KO mice exhibited low plasma LH concentrations (n�7–10 per group). Scale bar, 20�m.*p 0.05 versus controls.

7148 • J. Neurosci., July 6, 2016 • 36(27):7142–7153 McEwen et al. • SOCS3 Deletion Rescues from Leptin-Resistant Infertility

Figure 6. Neuron-specific SOCS3 knock-out delays the onset of HCD-induced leptin resistance in hypothalamic regions in female mice. Graphs represent quantification of pSTAT3-immunoreactivecell numbers in female mice (n �5–10 per group) perfused 2 h following a leptin injection (1 mg/kg). Mice were maintained on an HCD for either 110 d (A–E, sufficient to induce infertility in controlsbut not knock-outs) or 225 d (F–J, sufficient to induce infertility in both controls and knock-outs). K–O, Representative examples of leptin-induced pSTAT3 immunoreactivity in hypothalamic regionsin neuron-specific SOCS3 KO mice fed an HCD for 110 d (left images) or 225 d (right images). Scale bar, 200 �m. *p 0.05 versus LCD.

McEwen et al. • SOCS3 Deletion Rescues from Leptin-Resistant Infertility J. Neurosci., July 6, 2016 • 36(27):7142–7153 • 7149

The role of SOCS3 in HCD-induced obesity has been demon-strated using brain-specific SOCS3 knock-out mice (Mori et al.,2004). Supporting the findings of that study, neuron-specificSOCS3 knock-out completely corrected the overweight pheno-type exhibited by control male mice fed an HCD, although inter-estingly the protective effect of SOCS3 knock-out on the bodyweight and metabolic phenotype of females was considerablyless pronounced than in males in our experiments, perhaps re-flecting a difference in the mouse background strains used in thetwo studies or in the cell types targeted for Socs3 deletion(CamKinaseII� vs nestin-expressing cells). It is interesting tonote that even neuron-specific SOCS3 knock-out HCD-fed miceeventually developed resistance to leptin (defined as loss ofleptin-induced STAT3 activation) in most hypothalamic regions.This suggests that mechanisms other than SOCS3 upregulationcan eventually bring about hypothalamic leptin resistance. Thekey novel finding from this study is that female fertility when fed

on an HCD is improved by the absence of SOCS3 in diet-inducedinfertility-prone mice. This was despite the relatively limited ef-fects of SOCS3 knock-out in female metabolic function; there-fore, their improved fertility does not appear to be secondary toreduced obesity. Female neuron-specific SOCS3 knock-out micedid not become infertile at the same time as their control coun-terparts, and this was associated with an absence of leptin resis-tance onset after 110 d of HCD-feeding in the mPOA and AVPV,where susceptibility to leptin resistance has not previously beenreported (Munzberg et al., 2004). By comparison, control HCD-fed mice exhibited infertility and pronounced leptin resistance inthese regions. The rodent AVPV contains neurons that are acti-vated by the rising circulating estradiol concentration on the af-ternoon of proestrus (Le et al., 1997; Smith et al., 2006). Leptinacts on a small number of cells in this nucleus and to a muchgreater extent in the surrounding medial preoptic area (Fig. 7)(Quennell et al., 2009; Scott et al., 2009). Collectively, these re-sults may point to HCD-induced infertility in female mice beingcaused by a failure of ovulation, which is in turn due to SOCS3-induced onset of leptin resistance in the AVPV. Such site speci-ficity of the effects of leptin resistance would explain the lack ofcorrelation between blood SOCS3 concentrations and infertilityin women in a recent study (Ozkan et al., 2014).

To further examine this hypothesis, we tested whether thesuppressive effects of an HCD on estradiol-induced GnRH/LHsurge suppression were overcome by neuron-specific SOCS3knock-out. Although only 20% of control mice fed an HCDexperienced a preovulatory-like LH surge, LH surges could beinduced in almost all neuron-specific SOCS3 knock-out miceon this diet. In contrast, after 225 d of HCD feeding, when allmice in the fecundity test had become infertile, the ability ofestradiol to induce an LH surge was suppressed in both controland SOCS3 KO mice. It is important to note that, because theAVPV kisspeptin neurons which drive the preovulatory GnRHsurge do not express leptin receptors or show leptin signalingresponses (Cravo et al., 2011; Donato et al., 2011b; Louis et al.,2011; Quennell et al., 2011), the effects of leptin may insteadbe relayed to GnRH and/or kisspeptin neurons by nearby lep-tin receptor-expressing cells, such as neuronal nitric oxide-producing neurons in the mPOA (Donato et al., 2010). Theseneurons have recently been shown to be dependent in part onleptin for neuronal nitric oxide synthase activation, which is inturn required for leptin’s stimulatory effects on LH release andthe progression from diestrus to proestrus cycle stages (Belle-fontaine et al., 2014).

Another recently proposed site of action for leptin’s effectson fertility is the PMV, which contains neurons that expressleptin receptors and project to GnRH soma (Rondini et al.,2004; Leshan et al., 2009; Louis et al., 2011) and to AVPVkisspeptin neurons and GnRH terminals in the median emi-nence (Donato et al., 2011b). Ablation of this nucleus dis-rupted estrous cyclicity in female rats (Donato et al., 2009).Selective PMV leptin receptor reexpression in leptin receptornull mice induced female puberty onset and some pregnanciesbut had no effect on males (Donato et al., 2011b). However,glutamatergic neurons, which are very highly expressed in thePMV, are not required as direct mediators of leptin actions onreproductive activity (Zuure et al., 2013). It was therefore veryinteresting to note that HCD-induced infertility occurreddespite the PMV retaining normal leptin responsivenessthroughout the study. Lack of leptin resistance in HCD-fedmale C57BL/6J mice has previously been reported for the PMV(Munzberg et al., 2004).

Figure 7. Leptin responsiveness of voluntary food intake and circulating LH concentra-tion is suppressed after 160 d of HCD feeding in female mice. A, Twelve hour fastedLCD-fed mice showed a robust anorexigenic response in 3 h food intake to leptin (1 mg/kgi.p.) compared with saline vehicle. This response was completely suppressed in bothcontrol and SOCS3 KO HFD-fed mice (n � 5–7 per group). B, Twelve hour fasted LCD-fedmice exhibited a stimulation of LH secretion 20 min following leptin treatment (1 mg/kgi.p.), and this response was also suppressed in both genotypes of HFD-fed mice. **p 0.01 versus LCD controls.

7150 • J. Neurosci., July 6, 2016 • 36(27):7142–7153 McEwen et al. • SOCS3 Deletion Rescues from Leptin-Resistant Infertility

Another region shown to be protected from HCD-inducedleptin resistance by SOCS3 knock-out was the ARC. Althoughthe effects of HCD on leptin signaling in this region were notas pronounced as those in the mPOA and AVPV at 110 d ofHCD feeding, its likely importance as a site of leptin regulationof reproductive activity should not be overlooked. It containsnumerous GABAergic neurons that have recently been shownto be required as conduits for leptin’s permissive regulation offertility in mice (Zuure et al., 2013; Zuure et al., 2016), amongwhich agouti-related peptide/neuropeptide Y (Horvath et al.,1997; Israel et al., 2012; Roa and Herbison, 2012; Wu et al.,2012) and galanin-like peptide (Jureus et al., 2000; Takatsu etal., 2001) neurons may be important participants.

In contrast to the females, HCD-fed male mice did not exhibitany decline in fertility compared with LCD-fed males in thisstudy, despite their markedly increased body weight. This isconsistent with other studies (Tortoriello et al., 2004), althoughHCD-induced sperm defects have been reported (Bakos et al.,2010; Ghanayem et al., 2010; Fernandez et al., 2011). The sexspecificity of HCD feeding on fertility lends further support to theidea that the AVPV and its neuroendocrine control of ovulationmay be the principal targets for leptin resistant infertility.

Together with other data showing region-specific upregula-tion of Socs3 mRNA in response to HCD feeding (Munzberg etal., 2004), our data indicate that hypothalamic leptin resistanceand consequent female infertility caused by an HCD may be ini-tially induced by upregulation of SOCS3. Socs3 gene knock-outconfers protection against these effects. However, other mecha-nisms are also involved in the generation of leptin resistance, andthese mechanisms are likely to have been the cause of the eventualinfertility exhibited by the neuron-specific SOCS3 knock-outmice after �5 months of HCD feeding. It is possible that otherSOCS family proteins are upregulated to compensate for the lackof SOCS3. For example, SOCS1 levels are also increased alongwith SOCS3 in HCD-induced obesity (Peiser et al., 2000; Emanu-elli et al., 2008). Knocking out SOCS1 does not protect against thedevelopment of obesity, however (Emanuelli et al., 2008),suggesting that SOCS3 is the major SOCS inhibitor of hypotha-lamic leptin signaling. HCD-induced long-form leptin receptordownregulation is also thought to contribute to leptin resistance,and this may provide an explanation for how site-specific leptinresistance can occur in the absence of SOCS3. This downregula-tion of leptin receptor in response to an HCD is apparent insome studies (Tortoriello et al., 2004; Wilsey and Scarpace,2004), but not others (Munzberg et al., 2004; Jørgensen et al.,2006). Another potential mediator of SOCS3-independent leptinresistance is protein tyrosine phosphatase 1B (PTP1B). PTP1Binhibits both insulin and leptin signaling by dephosphorylatingactivated insulin receptor and insulin receptor substrates as wellas JAK2 (Salmeen et al., 2000), and PTP1B knock-out mice havemarkedly reduced adiposity when fed an HCD compared withcontrols (Tsou et al., 2012). It would be interesting to testwhether PTP1B knock-out mice would exhibit complete rescueof fertility when fed an HCD, or whether combined knock-out ofboth SOCS3 and PTP1B would be required for this.

SOCS3 also negatively regulates numerous other cytokine sig-naling pathways, including proinflammatory cytokines. SOCS3 isupregulated in the hypothalamus in response to inflammationfollowing a high fat diet (Wu et al., 2014), and is important forrestraining inflammation and allowing optimal levels of protec-tive immune responses against infection (Carow and Rottenberg,2014). Because central inflammation may itself be a cause ratherthan a consequence of leptin resistance and obesity (Wang et al.,

2012), SOCS3 knock-out could have affected leptin responsive-ness by this mechanism in the current study. However, the direc-tion of the response we observed (higher leptin responsiveness inthe SOCS3 knock-out mice) is not consistent with this. It remainspossible, however, that inflammation was a cause of the eventualonset of hypothalamic leptin resistance and infertility after pro-longed high calorie diet exposure.

In conclusion, deleting SOCS3 from forebrain neurons preve-nts HCD-induced suppression of the estradiol-induced GnRH/LH surge and fertility in female mice, as well as preventing HCD-induced leptin resistance from occurring in the mPOA and AVPV.However, this protective effect of SOCS3 deletion is not permanent,and it is likely that other longer-term mechanisms eventually bringabout leptin resistance and consequent female infertility. Furtherwork needs to be undertaken to determine how the different mech-anisms that cause leptin resistance change between short- and long-term feeding of HCD, as well as determining how the PMV appearsto remain entirely responsive to leptin under high caloric dietaryconditions while the mPOA and AVPV become leptin resistant, andthe sex-specific consequences of this.

NotesSupplemental material for this article is available at http://hdl.handle.net/10523/6230. This material has not been peer reviewed.

ReferencesBakos HW, Mitchell M, Setchell BP Lane M (2011) The effect of paternal

diet-induced obesity on sperm function and fertilization in a mousemodel. Int J Androl 34:402– 410. CrossRef Medline

Barash IA, Cheung CC, Weigle DS, Ren H, Kabigting EB, Kuijper JL, CliftonDK, Steiner RA (1996) Leptin is a metabolic signal to the reproductivesystem. Endocrinology 137:3144 –3147.

Bellefontaine N, Chachlaki K, Parkash J, Vanacker C, Colledge W,D’Anglemont De Tassigny X, Garthwaite J, Bouret SG, Prevot V (2014)Leptin-dependent neuronal NO signaling in the preoptic hypothalamusfacilitates reproduction. J Clin Invest 124:2550 –2559. CrossRef Medline

Bjørbaek C, Elmquist JK, Frantz JD, Shoelson SE, Flier JS (1998) Identifica-tion of SOCS-3 as a potential mediator of central leptin resistance. MolCell 1:619 – 625. CrossRef Medline

Burgin KE, Waxham MN, Rickling S, Westgate SA, Mobley WC, Kelly PT(1990) In situ hybridization histochemistry of Ca 2�/calmodulin-dependent protein kinase in developing rat brain. J Neurosci 10:1788 –1798. Medline

Carow B, Rottenberg ME (2014) SOCS3, a major regulator of infection andinflammation. Front Immunol 5:58. CrossRef Medline

Casanova E, Fehsenfeld S, Mantamadiotis T, Lemberger T, Greiner E, StewartAF, Schutz G (2001) A CamKIIalpha iCre BAC allows brain-specificgene inactivation. Genesis 31:37– 42. CrossRef Medline

Chehab FF, Lim ME, Lu R (1996) Correction of the sterility defect in homozy-gous obese female mice by treatment with the human recombinant leptin.Nat Genet 12:318 –320.

Chua SC Jr, Chung WK, Wu-Peng XS, Zhang Y, Liu SM, Tartaglia L, Leibel RL(1996) Phenotypes of mouse diabetes and rat fatty due to mutations in theOB (leptin) receptor. Science 271:994 –996.

Cravo RM, Margatho LO, Osborne-Lawrence S, Donato J Jr, Atkin S, Book-out AL, Rovinsky S, Frazao R, Lee CE, Gautron L, Zigman JM, Elias CF(2011) Characterization of Kiss1 neurons using transgenic mouse mod-els. Neuroscience 173:37–56. CrossRef Medline

Croker BA, Krebs DL, Zhang JG, Wormald S, Willson TA, Stanley EG, RobbL, Greenhalgh CJ, Forster I, Clausen BE, Nicola NA, Metcalf D, Hilton DJ,Roberts AW, Alexander WS (2003) SOCS3 negatively regulates IL-6 sig-naling in vivo. Nat Immunol 4:540 –545. CrossRef Medline

Dokras A, Baredziak L, Blaine J, Syrop C, VanVoorhis BJ, Sparks A (2006)Obstetric outcomes after in vitro fertilization in obese and morbidly obesewomen. Obstet Gynecol 108:61– 69. CrossRef Medline

Donato J Jr, Silva RJ, Sita LV, Lee S, Lee C, Lacchini S, Bittencourt JC, FranciCR, Canteras NS, Elias CF (2009) The ventral premammillary nucleuslinks fasting-induced changes in leptin levels and coordinated luteinizinghormone secretion. J Neurosci 29:5240 –5250. CrossRef Medline

McEwen et al. • SOCS3 Deletion Rescues from Leptin-Resistant Infertility J. Neurosci., July 6, 2016 • 36(27):7142–7153 • 7151

Donato J Jr, Frazao R, Fukuda M, Vianna CR, Elias CF (2010) Leptininduces phosphorylation of neuronal nitric oxide synthase in defi-ned hypothalamic neurons. Endocrinology 151:5415–5427. CrossRefMedline

Donato J Jr, Cravo RM, Frazao R, Elias CF (2011a) Hypothalamic sites ofleptin action linking metabolism and reproduction. Neuroendocrinology93:9 –18. CrossRef Medline

Donato J Jr, Cravo RM, Frazao R, Gautron L, Scott MM, Lachey J, Castro IA,Margatho LO, Lee S, Lee C, Richardson JA, Friedman J, Chua S Jr, Cop-pari R, Zigman JM, Elmquist JK, Elias CF (2011b) Leptin’s effect onpuberty in mice is relayed by the ventral premammillary nucleus anddoes not require signaling in Kiss1 neurons. J Clin Invest 121:355–368.CrossRef Medline

Emanuelli B, Macotela Y, Boucher J, Ronald Kahn C (2008) SOCS-1 defi-ciency does not prevent diet-induced insulin resistance. Biochem BiophysRes Commun 377:447– 452. CrossRef Medline

Evans MC, Rizwan MZ,Anderson GM (2015) Insulin does not targetCamkII� neurons to critically regulate the neuroendocrine reproductiveaxis in mice. J Neuroendocrinol 27:899 –910. CrossRef Medline

Fernandez CD, Bellentani FF, Fernandes GS, Perobelli JE, Favareto AP, Nas-cimento AF, Cicogna AC, Kempinas WD (2011) Diet-induced obesityin rats leads to a decrease in sperm motility. Reprod Biol Endocrinol 9:32.CrossRef Medline

Franklin BJ, Paxinos G (2008) The mouse brain atlas in stereotaxic coordi-nates. New York: Elsevier.

Franz WB 3rd (1988) Basic review: endocrinology of the normal menstrualcycle. Prim Care 15:607– 616. Medline

Ghanayem BI, Bai R, Kissling GE, Travlos G, Hoffler U (2010) Diet-inducedobesity in male mice is associated with reduced fertility and potentiationof acrylamide-induced reproductive toxicity. Biol Reprod 82:96 –104.CrossRef Medline

Homburg R, Potashnik G, Lunenfeld B, Insler V (1976) The hypothalamusas a regulator of reproductive function. Obstet Gynecol Surv 31:455– 471.CrossRef Medline

Horvath TL, Bechmann I, Naftolin F, Kalra SP, Leranth C (1997) Heteroge-neity in the neuropeptide Y-containing neurons of the rat arcuate nu-cleus: GABAergic and non-GABAergic subpopulations. Brain Res756:283–286. CrossRef Medline

Howard JK, Flier JS (2006) Attenuation of leptin and insulin signaling bySOCS proteins. Trends Endocrinol Metab 17:365–371. CrossRef Medline

Israel DD, Sheffer-Babila S, de Luca C, Jo YH, Liu SM, Xia Q, Spergel DJ, DunSL, Dun NJ, Chua SC Jr (2012) Effects of leptin and melanocortin sig-naling interactions on pubertal development and reproduction. Endocri-nology 153:2408 –2419. CrossRef Medline

Jørgensen EA, Vogelsang TW, Knigge U, Watanabe T, Warberg J, Kjaer A(2006) Increased susceptibility to diet-induced obesity in histamine-deficient mice. Neuroendocrinology 83:289 –294. CrossRef Medline

Jureus A, Cunningham MJ, McClain ME, Clifton DK, Steiner RA (2000)Galanin-like peptide (GALP) is a target for regulation by leptin in thehypothalamus of the rat. Endocrinology 141:2703–2706. CrossRefMedline

Koning AM, Kuchenbecker WK, Groen H, Hoek A, Land JA, Khan KS, MolBW (2010) Economic consequences of overweight and obesity in infer-tility: a framework for evaluating the costs and outcomes of fertility care.Hum Reprod Update 16:246 –254. CrossRef Medline

Le WW, Attardi B, Berghorn KA, Blaustein J, Hoffman GE (1997) Pro-gesterone blockade of a luteinizing hormone surge blocks luteinizinghormone-releasing hormone Fos activation and activation of its preopticarea afferents. Brain Res 778:272–280. CrossRef Medline

Le WW, Wise PM, Murphy AZ, Coolen LM, Hoffman GE (2001) Paralleldeclines in Fos activation of the medial anteroventral periventricularnucleus and LHRH neurons in middle-aged rats. Endocrinology 142:4976 – 4982. CrossRef Medline

Lebrethon MC, Vandersmissen E, Gerard A, Parent AS, Bourguignon JP(2000) Cocaine and amphetamine-regulated-transcript peptide media-tion of leptin stimulatory effect on the rat gonadotropin-releasing hor-mone pulse generator in vitro. J Neuroendocrinol 12:383–385. Medline

Leshan RL, Louis GW, Jo YH, Rhodes CJ, Munzberg H, Myers MG Jr (2009)Direct innervation of GnRH neurons by metabolic- and sexual odorant-sensing leptin receptor neurons in the hypothalamic ventral premammil-lary nucleus. J Neurosci 29:3138 –3147. CrossRef Medline

Louis GW, Greenwald-Yarnell M, Phillips R, Coolen LM, Lehman MN, My-

ers MG Jr (2011) Molecular mapping of the neural pathways lin-king leptin to the neuroendocrine reproductive axis. Endocrinology 152:2302–2310. CrossRef Medline

Mori H, Hanada R, Hanada T, Aki D, Mashima R, Nishinakamura H, TorisuT, Chien KR, Yasukawa H, Yoshimura A (2004) Socs3 deficiency in thebrain elevates leptin sensitivity and confers resistance to diet-inducedobesity. Nat Med 10:739 –743. CrossRef Medline

Munzberg H, Flier JS, Bjørbaek C (2004) Region-specific leptin resistancewithin the hypothalamus of diet-induced obese mice. Endocrinology 145:4880 – 4889. CrossRef Medline

Nagatani S, Zeng Y, Keisler DH, Foster DL, Jaffe CA (2000) Leptin regulatespulsatile luteinizing hormone and growth hormone secretion in thesheep. Endocrinology 141:3965–3975. CrossRef Medline

Ottaway N, Mahbod P, Rivero B, Norman LA, Gertler A, D’Alessio DA,Perez-Tilve D (2015) Diet-induced obese mice retain endogenous leptinaction. Cell Metab 21:877– 882. CrossRef Medline

Ouimet CC, McGuinness TL, Greengard P (1984) Immunocytochemicallocalization of calcium/calmodulin-dependent protein kinase II in ratbrain. Proc Natl Acad Sci U S A 81:5604 –5608. CrossRef Medline

Ozkan ZS, Deveci D, Kumbak B, Simsek M, Ilhan F, Sekercioglu S, SapmazE (2014) What is the impact of Th1/Th2 ratio, SOCS3, IL17, andIL35 levels in unexplained infertility? J Reprod Immunol 103:53–58.CrossRef Medline

Peiser C, McGregor GP, Lang RE (2000) Leptin receptor expression andsuppressor of cytokine signaling transcript levels in high-fat-fed rats. LifeSci 67:2971–2981. CrossRef Medline

Quennell JH, Mulligan AC, Tups A, Liu X, Phipps SJ, Kemp CJ, HerbisonAE, Grattan DR, Anderson GM (2009) Leptin indirectly regulatesgonadotropin-releasing hormone neuronal function. Endocrinology150:2805–2812. CrossRef Medline

Quennell JH, Howell CS, Roa J, Augustine RA, Grattan DR, Anderson GM(2011) Leptin deficiency and diet-induced obesity reduce hypothalamickisspeptin expression in mice. Endocrinology 152:1541–1550. CrossRefMedline

Rizwan MZ, Poling MC, Corr M, Cornes PA, Augustine RA, QuennellJH, Kauffman AS, Anderson GM (2012) RFamide-related peptide-3receptor gene expression on GnRH and kisspeptin neurons and GnRH-dependent mechanism of action. Endocrinology 153:3770 –3779.CrossRef Medline

Roa J, Herbison AE (2012) Direct regulation of GnRH neuron excitabilityby arcuate nucleus POMC and NPY neuron neuropeptides in femalemice. Endocrinology 153:5587–5599. CrossRef Medline

Rondini TA, Baddini SP, Sousa LF, Bittencourt JC, Elias CF (2004) Hy-pothalamic cocaine- and amphetamine-regulated transcript neuronsproject to areas expressing gonadotropin releasing hormone immuno-reactivity and to the anteroventral periventricular nucleus in male andfemale rats. Neuroscience 125:735–748. CrossRef Medline

Salmeen A, Andersen JN, Myers MP, Tonks NK, Barford D (2000) Molec-ular basis for the dephosphorylation of the activation segment of theinsulin receptor by protein tyrosine phosphatase 1B. Mol Cell 6:1401–1412. CrossRef Medline

Scott MM, Lachey JL, Sternson SM, Lee CE, Elias CF, Friedman JM, ElmquistJK (2009) Leptin targets in the mouse brain. J Comp Neurol 514:518 –532. CrossRef Medline

Smith JT, Popa SM, Clifton DK, Hoffman GE, Steiner RA (2006) Kiss1 neu-rons in the forebrain as central processors for generating the preovulatoryluteinizing hormone surge. J Neurosci 26:6687– 6694. CrossRef Medline

Souter I, Baltagi LM, Kuleta D, Meeker JD, Petrozza JC (2011) Women,weight, and fertility: the effect of body mass index on the outcome ofsuperovulation/intrauterine insemination cycles. Fertil Steril 95:1042–1047. CrossRef Medline

Takatsu Y, Matsumoto H, Ohtaki T, Kumano S, Kitada C, Onda H,Nishimura O, Fujino M (2001) Distribution of galanin-like peptide inthe rat brain. Endocrinology 142:1626 –1634. CrossRef Medline

Tortoriello DV, McMinn J, Chua SC (2004) Dietary-induced obesity andhypothalamic infertility in female DBA/2J mice. Endocrinology 145:1238 –1247. CrossRef Medline

Tsou RC, Zimmer DJ, De Jonghe BC, Bence KK (2012) Deficiency of PTP1Bin leptin receptor-expressing neurons leads to decreased body weight andadiposity in mice. Endocrinology 153:4227– 4237. CrossRef Medline

7152 • J. Neurosci., July 6, 2016 • 36(27):7142–7153 McEwen et al. • SOCS3 Deletion Rescues from Leptin-Resistant Infertility

van Dielen FM, van’T Veer C, Buurman WA, Greve JW (2002) Leptin andsoluble leptin receptor levels in obese and weight-losing individuals.J Clin Endocrinol Metab 87:1708 –1716. CrossRef Medline

Wang X, Ge A, Cheng M, Guo F, Zhao M, Zhou X, Liu L, Yang N (2012)Increased hypothalamic inflammation associated with the susceptibilityto obesity in rats exposed to high-fat diet. Exp Diabetes Res 2012:847246.CrossRef Medline

Wauman J, Tavernier J (2011) Leptin receptor signaling: pathways toleptin resistance. Front Biosci (Landmark Ed) 17:2771–2793. CrossRefMedline

Wilsey J, Scarpace PJ (2004) Caloric restriction reverses the deficits in leptinreceptor protein and leptin signaling capacity associated with diet-induced obesity: role of leptin in the regulation of hypothalamic long-form leptin receptor expression. J Endocrinol 181:297–306. CrossRefMedline

Wintermantel TM, Campbell RE, Porteous R, Bock D, Grone HJ, TodmanMG, Korach KS, Greiner E, Perez CA, Schutz G, Herbison AE (2006)

Definition of estrogen receptor pathway critical for estrogen positive feed-back to gonadotropin-releasing hormone neurons and fertility. Neuron52:271–280. CrossRef Medline

Wu Q, Whiddon BB, Palmiter RD (2012) Ablation of neurons expressingagouti-related protein, but not melanin concentrating hormone, inleptin-deficient mice restores metabolic functions and fertility. Proc NatlAcad Sci U S A 109:3155–3160. CrossRef Medline

Wu Y, Yu Y, Szabo A, Han M, Huang XF (2014) Central inflammationand leptin resistance are attenuated by ginsenoside Rb1 treatment inobese mice fed a high-fat diet. PLoS One 9:e92618. CrossRef Medline

Zuure WA, Roberts AL, Quennell JH, Anderson GM (2013) Leptinsignaling in GABA neurons, but not glutamate neurons, is requiredfor reproductive function. J Neurosci 33:17874 –17883. CrossRefMedline

Zuure WA, Quennell JH, Anderson GM (2016) Leptin responsive andGABAergic projections to the rostral preoptic area in mice. J Neuroendo-crinol 2016:3. CrossRef Medline

McEwen et al. • SOCS3 Deletion Rescues from Leptin-Resistant Infertility J. Neurosci., July 6, 2016 • 36(27):7142–7153 • 7153